March, 1948
CARCINOGENIC CHOLANTHRENE HYDROCARBONS
1073
Summary pared by an improved synthesis for benz [alanthra1. 9,ll-Dirnethylbenz [alanthracene and 8,9,- cene derivatives. 11-trimethylbenz [a]anthracene have been pre- EVANSTON, ILLINOIS RECEIVED AUGUST15, 1947
[CONTRIBUTION FROM THE CHEMICAL
LABORATORY O F NORTHWESTERN UNIVERSITY]
Carcinogenic Hydrocarbons. 3,5-Dimethylcholanthrene1 BY BYRONRIEGEL,JOHN G. B U R R ,
JR.,? ~ ~ I C I I A F XA. ,
KUBICO AND MARVIN 15. GOLD3
Since the preparation of methylcholanthrene this substance conform well with those expected from desoxycholic acid," and from cholic acid5 by for l-keto-3,5-dimethylcholanthrene.It is evia series of reactions whose types are known to oc- dent that extensive dehydrogenation has occurred, cur in the animal body, it has been a natural specu- presumably by hydrogen transfer during the cylationfithat some tumors may be caused by car- clization, although the low hydrogen content found cinogenic agents produced by an abnormal metab- in the acid indicates the possibility that dehydroolism of sterols or bile acids. The formation of genation occurred in an earlier step. At the time substituted cholanthrenes from steroids requires a this experimental work was performed, the extent ring closure between the steroid side chain and of the dehydrogenation was not suspected, and ring C of the nucleus. The residue of the side both the non-crystalline and the crystalline cychain will then occupy the 5 position in cholan- clization products were combined, after reduction threne, and the C-20 methyl group of the steroid of the keto group, and carried through to 3,5-diwill be at the ;3 position of cholanthrene (present methylcholanthrene which was obtained in fair system of numbering7). To date, 3,S-dialkyl- yield from the mixture. cholanthrenes have not been prepared and tested The reduction of the ketonic material was acfor carcinogenic activity. The biological reasons complished by the Martin modification of the for desiring such information are obvious. Clemmensen reduction. lo The reduction prodThe first member of this series, 3,5-dimethyl- ucts from both the solid and the oily ketone were cholanthrene (HI), has been prepared by general combined. Two methods were used for aromatimethods already recorded in the literature8 for zation of this material: (1) the Linstead liquid the preparation of substituted cholanthrenes. phase method" which gave pure 3,5-dimethylThe starting material was 8-keto-9,1l-dimethyl- cholanthrene (111) but in poor yield, and ( 2 ) the 8,9,10,11-tetrahydrobenz [alanthracene (I).9 By heating of the crude tetrahydro hydrocarbon with means of a Reformatsky reaction followed by de- palladiurn-charcoal at 300". This last method hydration, hydrogenation, and saponification, the was distinctly superior, giving the 3,5-dimethylketone (I) was transformed into 9,ll-dimethyl- cholanthrene (111) in a 51% yield. 8,9,10,11-tetrahydrobenz [a]anthryl-8-acetic acid (11). Cyclization of this acid by the action of stannic chloride upon the acid chloride produced a viscous liquid ketone which after reduction and dehydrogenation gave ~~,5-dimethylcholanthrene. The action of liquid hydrogen fluoride upon the acid gave two products, one of which was an acetone-soluble viscous liquid, and the other (35$%) 1 was an acetone-insoluble substance crystallizing in bright yellow needles with a melting point of 2672 i O o . The physical properties and analysis o f ( 1 ) From the I'h. n. Theses by Michael A. Kubico, l!M3 aiirl J o h n G. Burr, J r , 1!147. ( 2 ) Eastm.m Lodirk Fellow, 1946-1947. Present address: Department of Chemistry, Miami University, Oxford, Ohio. (a) Anna Fuller F u n d Research Associate, 1940-1942. (4) Ii. Wieland and 17. Dane, Z . pizysiol. Chcm., 219, 240 (1933). i.5) I,. 1;. r'ieser and AI. S. Newman, T H i s JOURNAL, 67, 961 (19351 ( 6 ) J. W. Cook and G. A. D. Haslewood, Chemistry and I7rduslry. 52,758 (19331; J. Chem. SOC.,428 (1934). (7) A. M. Patf.erson and L. T. Capell, "The Ring Index," Reinhold Publishing Corp., New York, N. Y., 1940, p. 437. (8) (a) J. W. Cook and G. A. D. Haslewood, J. Chem. Soc., 767 (1935); (b) E. Beryman and 0.Blum-Bergmann, THIS JOURNAL, 69, 1573 (1937); ( c ) !A.' E. Bachmann, J. Orp. Chem., 3, 434 (1938); id) W. E. Bachmann and J. M. Chemerda, ibid., 6 , 50 (1941). (9) B. Riegel and .I G. Burr, J r . , THISJ O U R N A L , T O , 1070 (1948).
111
1. .!
From ethanol or ethanol-acetone 3,5-dimethylcholanthrene crystallized as orange-yellow needles melting at 183-184", but from acetone alone, it crystallized either as diamond-shaped plates or (10) E. L. Martin, i b i d . , 68, 1438 (1936). (11) R. P. Linstead, A. F. Millidge, S. I,. S . Thomas and A. L. Walpole, J . Chpm. S ~ I C -1.l,l 6 (1937).
1074
BYRONRIEGEL,JOHN G. BURR,JR., MICHAEL A. KUBICOAND MARVINH. GOLD
stout rods all having the same melting point alone or admixed. An acetone or alcohol solution of the hydrocarbon has a strong violet fluorescence. When 3,5-dimethylcholanthreneis dissolved in concentrated sulfuric acid, a rose colored solution with a yellow fluorescenceis formed. The hydro-
Vol. 70
4-chloro-1,3-dimethylbenzene(V), would give the
same and desired hydrindene (VII) after cyclization and Clemmensen reduction of the chlorodimethylhydrindones. One of these isomers, tentatively assigned the structure written as VI, was obtained as a sharp melting crystalline product. I t gave a high yield (94%) of chlorodimethylhydrinTABLEI done on ring closure with concentrated sulfuric MAXIMAIN THE ULTRAYIOLET ABSORPTION SPECTRAOF acid. The chlorine atom in VI1 proved so highly CHOLANTHRENES (X IN A. UNITS) AND CORRESPONDING hindered that a negligible yield of VI11 was obINTENSITIES (LOGEM); SOLVENT, ETHANOL tained by condensation of the lithium derivative Cholanthrene with a-naphthonitrile. Attempts to exchange the and dimethyl derivatives A B c D E F chlorine atom in 4-chloro-5,7-dimethylhydrindene Cholanthrene" 2615 2740 2845 2950 3280 3420 (VI11 for a cyano g o u p and then condense the rewith a4 . 6 4 . 6 4 . 8 4 . 9 3 . 6 3 . 5 sulting 4-cyano-b,7-dimethylhydrindene 4,1 l-Diinethyl'2 2635 2780 2882 2985 3310 3445 naphthylmagnesium bromide were likewise fruit4.64 4.63 4.78 4.82 3.69 3.80 less. 3,11-Dirnethyl'2 2610 2768 2868 2983 3255 3430 An alternative path was sought in the reaction 4.65 4 . 6 5 4.89 4.98 3.69 3 . 5 5 of 3'-keto-3,4-dihydro-1,2-cyclopentenophenan3,5-Dimethyl 2650 2770 2880 3000 3300 3460 threne (Ix)with ethyl a-bromopropionate. It 4 . 6 0 4.55 4.81 4.94 3.63 3.80 was planned to extend the side chain of the reby the Arndt-Eisted proceo L. F. Fieser and E. E. Hershberg, THISJOURNAL, 60, sulting acid dure and then by aromatization and cyclization 940 (1935). to arrive at the ketone (XI). 3,3-Dimethylcholcarbon was characterized12by means of its ultra- anthrene could be derived from this ketone (XI) violet absorption spectrum. The curve for this by an obvious series of reactions. However, the exhibits the maxima characteristic of the benz[a]- reaction of the ketone (IX) with ethyl a-bromoanthracene system which show a bathochromic propionate gave a product which could not be shift of 3 0 4 0 A. units when compared to positions characterized either as the ester or as the free acid. of the maxima of cholanthrene (Table I). The Several other synthetic schemes were also exsame shift has been observed for the two methyl groups in 3,11CH2 CH3 and 4,11-dimethylcholanthrene l I (Table I). The method which has been -2-+ most widely used for the preparaI tion of substituted cholanthrenes I c1 c1 is the Elbs pyrolysis of the appropriately substituted naphthoylhy1VI drinder~es.'~In spite of the fact that pyrolysis of the ketone (VIII) could yield two possible isomers, caused by ring closure to either the o-methyl group or the o-methylene group, several efforts were made to \ prepare this substance with the VI11 aim of pyrolyzing it to 3,S-diCOOH methylcholanthrene (111). By the I chlorination of xylene,l 4 4-chloro0 CHCHa 1,3-dimethylbenzene (V) was pre-. pared. B Friedel-Crafts condensation of this compound with /3chloropropionyl chloride should produce theoretically three different + isomeric chlorodimethyl-B-chloropropiophenones. Two of these posIX x XI sible isomers, i. e., where substitution has taken place a t the 5 or 6 positions of the plored but with less success than those described. Eamples of 3,3-dimethylcholanthrenehave been (12) L. 17. Fieser and D. M. Bowen, TIirs JOURNAL, 62, 2103 submitted to Dr. M. J. Shear, National Cancer (1940). Institute, and Dr. A. M. Seligman, for biological (13) L. F. Fieser and A . M. Seligman, ibid., 68, 2482 (1936). testing. ( 1 4 ) 0. Jacobsen, Ber., 18, 17Fl (1885).
CARCINOGENIC CHOLANTHRENE HYDROCARBONS
March, 1948
Experimental16 9,11-Dimethyl-8,9,10,11 -tetrahydrobenz[alanthryl-8acetic Acid (II).-On warming a mixture of 2.0 g. of 8keto - 9,11 dimethyl - 8,9,10,11 tetrahydrobenz[a]a n t h r a ~ e n e3, ~ml. of ethyl a-bromoacetate, 0.2 g. of iodine, 4.0 g. of 20-mesh zinc (thoroughly cleaned and dried), 20 ml. of dry ether and 20 ml. of dry benzene, a brisk spontaneous reaction set in. When the reaction subsided (fifteen minutes), the solution was refluxed for three hours, and hydrolyzed with dilute hydrochloric acid. The aqueous layer was extracted three times with ether, then the combined organic solutions were dried and concentrated. The deep red residue was warmed gently in wacuo and dissolved in a mixture of 30 ml. of dry alcohol and 20 ml. of dry benzene containing some anhydrous sodium sulfate. Dry hydrogen chloride was passed into the solution (chilled in an ice-bath) for one-half hour. The chloro compound was recovered by diluting the solution with water and extracting with ether. The ether solution was washed with dilute potassium carbonate solution, dried, and the solvent was removed. Dehydrohalogenation was effected by distilling the dark residue under two microns pressure at a bath temperature of 140-190". The distillation was carried out very slowly to minimize a persistent tendency to bump. The light amber viscous distillate weighed 2.43 g. (98yo). On one larger run, the distillation was stopped when 14.0 g. (78%) of the unsaturated ester had distilled during a period of one-hundred forty-four hours at lo-' mm. (bath a t 115-120 ") In this case the still residue was a dark stiff glass which weighed 5.46 g. (31%). The unsaturated ester (2.43 g., 0.0071 mole) was hydrogenated in absolute ethanol, using 0.34 g. of platinum oxide. During one and three-quarters hours, it absorbed 136 ml. of hydrogen at 27" (78% of the theoretical). This same hydrogen uptake was observed in a number of runs which varied in size. The saturated ester was obtained as a viscous yellow liquid. It was saponified in a nitrogen atmosphere with alcoholic potassium hydroxide. The alkaline solution was diluted with mater, extracted with ether and acidified. The acid (11) precipitated as light yellow flakes and crystals which weighed 2.04 g. (87%) after filtering and air-drying. In other runs, yields as high as 9570 were observed, based on the dimethyl ketone (I). A portion of the acid was crystallized once from ethanol, then from methanol, and W ~ obtained S as colorless needles which melted a t 192.5-194 . Anal. Calcd. for CzzHzz02: C, 83.0; H, 6.96. Found: C, 83.3; 11, 6.30. l-Keto-3,5-dirnethylcholanthrene.-The acid (11), 11.36 g. (including 3.06 g. of non-crystalline material), was cyclized by dissolving it in about 150 g. of anhydrous liquid hydrogen fluoride contained in a copper beaker surrounded by an ice-bath. After evaporation of the hydrogen fluoride, the residue was washed with water and with a solution of potassium carbonate, then triturated with acetone. The resulting solid was suspended in boiling benzene which was cooled slightly and diluted with ether. The solid product (yellow-brown crystals) was separated by filtration and crystallized twice from chloroform and twice from benzene. 11. was very difficultly soluble in both solvents, as well as acetone. The ketone forms brilliant yellow needles mcslting a t 267-270°, and it gives a precipitate when heated with alcoholic 2,4-dinitrophenylhydrazine. Anal. Calcd. for C22H160: C, 89.2; H, 5.40. Found: c, 89.6; E l , 5.57 The oxime crystallized from chloroform as fluffy yellow needles which darkened a t 245" and decomposed without melting a t 254-258O. A n d . Calcd. for C ~ ~ H ~ T ON, N :4.50. Found: N, 4.42. A total of 4.7 g. of this solid ketone was obtained. The acetone-soluble fraction composed 8.75 g. of a stiff red-
-
-
.
(15) All meltlug p u i u t s were taken on a Fischer-Johns apparatus. Ilicroanalyses are by Patricia Craig and Nelda Mold, Northwestera University.
1075
brown oil. This was considered to be the tetrahydro ketone. The solid ketone (3.5 g.), and the oily ketone (8.75 8.) were dissolved separately in 34-ml. portions of glacial acetic acid. To each of these solutions was added 50 ml. of toluene, 34 ml. of water, 200 ml. of concentrated hydrochloric acid, and 65 g. of amalgamated mossy zinc. The solutions were refluxed for forty-eight hours; 150 ml. of concentrated hydrochloric acid was added in 50-ml. portions at twelve-hour intervals. From each of these experiments, when worked up, was obtained a deep red gum. During the work-up, 1.3 g. of acidic material was recovered, presumably uncyclized (11). The products were combined, giving 10.1 g. of hydrocarbon. 3,5-Dimethylcholanthrene (III). A.-A 2-g. sample of the above product was dissolved in 50 ml. of @-cymene (terpene-free) , 0.15 g. of 5% palladium-charcoal added and the solution heated t o brisk boiling in a stream of carbon dioxide. The gases were collected in a nitrometer. In all, 202 ml. of inert gas (uncorrected for blank) were obtained during several days. The product recovered from this solution was an amber viscous substance (1.46 g.) which could not be made t o crystallize. It was distilled under 4 microns pressure (bath at 140-170'). The 0.90 g. of light orange distillate (a viscous liquid) was crystallized from ethanol-benzene, giving 0.30 g. of orangeyellow needles, m. p. 170-173 O . 3,5-Dimethylcholanthrene, after three crystallizations from e t h a n o l a c e t y e , formed pale orange-yellow needles melting at 183-184 , Anal. Calcd. for CnHls: C, 93.6; H, 6.45. Found: C, 93.4; H, 6.49. B. Another sample of the Clemmensen product, 10.1 g., was heated with 1.50 g. of 5% palladium-charcoal a t 300-310" for fortyominutes. A brisk evolution of gas occurred at 285-310 The cooled melt was repeatedly extracted with benzene; and the benzene solution, after filtering, was concentrated. The solution was diluted with a n equal volume of ethanol, decolorized, and filtered. The filtrate gave 2.2 g. of yellow-brown plates, m. p. 175179'. The mother liquors were evaporated to dryness and the residue triturated with boiling ethanol. An additional 2.7 g. of crystalline hydrocarbon was thus obtained. The ethanol solution deposited 0.5 g. of hydrocarbon, making a total yield of 5.5 g. (51%). The crude hydrocarbon was decolorized in acetone solution, then a benzene solution of i t passed through alumina. The material recovered from the eluate was crystallized from acetone. A total of 3.28 g. of shining yell;w-brown plates was ohtajned, which melted a t 179-181 A mixture of this material with the needle form of the hydrocarbon showed no depression of the melting point; it crystallized from ethanol as needles. 3,5-Dimethylcholanthrene forms a picrate, m. p. 179180 which dissociates when recrvstallized from ethanol. Anal. Calcd. for C22H~C~Hafi307:N, 8.22. Found: N, 8.22. The trinitrobenzene derivative forms mas$ve dull red needles from ethanol, which melt a t 189-190 . Anal. Calcd. for C22Hj8.CsH~NaOe: N, 8.48. Found: N, 8.24. The trinitrofluorenone derivaTe crystallizes as blue microneedles, melting at 252-255 A Beckmann spectrophotometer was used for observation of the ultraviolet spectrum. The hydrocarbon, weighed on a microbalance, was dissolved in absolute ethanol, and the solution diluted accurately to a concentration of 0.0025 mg./ml. p-(a-Naphthoyl) -propionic Acid.-The intermediate used for the preparation of compound (IX) was l-keto1,2,3,4-tetrahydrophenanthrene. Several methods were investigated for the preparation of this intermediate because it is a laborious task. One of the possible starting materials for its synthesis is p-(a-naphthoyl) -propionic acid. The following new method for the preparation of this starting material was explored. A benzene solution of di-a-naphthylcadmium (prepared by the general method
.
.
".
.
1076
BYRONRIEGEL,JOHN G. BURR,JR., MICHAEL A. KUBICO AND MARVIN H.GOLD
of Cason and Prout" from 20.7 g. of a-bromonaphthalene and 9.7 g. of anhydrous cadmium chloride) reacted with 16.5 g. of carbomethoxypropionyl chloride. The crude ester isolated' from this reaction was a pale yellow oil boiling at 184-188'' (1 mm.). It was saponified with alcoholic potassium hydroxide, giving 10.2 g. (45%) of the free acid which after one crystallization from dilute ethanol melted at 133-134' (lit." 132-133'). This acid was also prepared by the customary17 acylation of naphthalene with succinic anhydride, using tetrachloroethane as solvent in place of nitrobenzene. Reaction between 3'-Keto-3,4-dihydro-1,2-cyclopentenophenanthrene (IX) 18 and Ethyl a-Bromopropionate.A mixture of 2.20 g. of the ketone (X) (m. p . 214-216"), 2.09 g. of ethyl a-bromopropionate, 5 g. of thoroughly cleaned and dried 20-mesh zinc, 0.10 g. of iodine, 70 ml. of benzene, and 30 ml. of toluene was refluxed for four and one-half hours. The iodine color faded immediately, and the pale yellow solution became orange as the reaction progressed. Two additions of 2.5 g. of zinc were made at forty-five minute intervals, and 1.90 g. of bromoester was added after one and one-half hours. The neutral material isolated from this reaction was distilled in high vacuum; and the distillate, after numerous fruitless attempts at crystallization, was saponified with alcoholic potassium hydroxide. The acidic material so obtained was 2.52 g. (93%) of a very viscous liquid, which could not be made to crystallize. Only a small fraction of it was volatile in a high vacuum. When this reaction was carried out in etherbenzene solution, the ketone (IX) was recovered almost quantitatively. Chloro-m-dimethyl- 0-chloropropiophenone (VI) Aluminum chloride, 72 g., was added in portions t o a solution of 36 g. of ,khloropropionyl chloride and 40 g. of 4-chloro-l,34imethylbenzene14 (b. p. 187-194' (760 mm.)) in 145 ml. of carbon disulfide. The addition required two t o three hours. Stirring was continued overnight, and the reaction then heated briefly to 45'. The mixture was hydrolyzed, and the solvent layer after drying was distilled (keeping the temperature below 60 ') The residue crystallized on chilling; it weighed 67.5 g. Trituration of this solid with methanol gave 20.7 g. of slightly yellow solid, m. p. 84-87', suitabIe for cyclization. Seven crystallizations of this solid from petroleum ether produce: the p-chloroketone (VI) with a melting point of 93-93.5 Anal. Calcd. for CllH120C12: C, 57.1; H , 5.24. Found: C, 57.2; H, 5.32. x-Chloro-m-dimethy1hydrindone.-A solution of 19 g. of the crude crystalline p-chloroketone (VI) in 50 ml. of benzene was slowly run into 150 ml. of concentrated sulfuric acid a t 105-120°, with rapid stirring. The mixture was stirred for twenty minutes, then poured onto cracked
.-
.
.
(16) J. Cason and F. S. Prout, THIS J O U R N A L , 66, 46 (1944). (17) L. F. Fieser and M. A. Peters, ibid., 64, 4347 (1932). (18) Prepared by the method of W. S. Johnson and J. W. Petersen, i b i d . , 67, 1.366 (1945).
Vol. 70
ice. The precipitated solid, 15 g. (93.5%), when recrystallized once from methanol melted a t 110-111 o. The non-crystalline substance recovered from the methanol washings above was cyclized separately. An additional 8 g. of hydrindone was so obtained. Anal. Calcd. for CllHllOCl: C, 67.8; H, 5.70. Found: C,67.8; H, 5.77. 4-Chloro-5,7-dimethylhydrindene(VII) .-The chlorohydrindone, 10 g., was reduced by the Martin modification of the Clemmensen procedure.lO The h drindene was isolated as a colorless liquid, 8.67 g. (94&, which boiled a t 126-129" (12 mm.): d26 1.116: n 1 9 ~1.5549. Anal. 'Calcd. fo; CIIH13C1:'C,73.1; H, 7.25. Found: C, 73.0; H, 7.15. Attempts to Condense the Hydrindene (VII) with (YNaphthonitri1e.-A mixture of 3.77 g. of the hydrindene (VII), 30 ml. of absolute ether and 0.32 g. of lithium metal cut into tiny pieces was refluxed and stirred in a nitrogen atmosphere for twenty-four hours. Although some of the lithium remained unreacted, 3.19 g. of a-naphthonitrile dissolved in 40 ml. of dry benzene was added and reflux with stirring continued for three hours. The material obtained from hydrolysis of the reaction mixture was refluxed with alcoholic hydrochloric acid, then distilled under reduced pressure. A large amount of the hydrindene (VII) was recovered and a very small amount of high boiling material. When di-n-butyl ether was used as the solvent for this reaction, 2.0 g. of the hydrindene required nineteen hours to react with 0.19 g. of lithium metal. However, treatment of the solution with 1.69 g. of the nitrile resulted in almost quantitative recovery of both starting materials.
Summary
1. 3,5-Dimethylcholanthrene has been prepared from 8-keto-9,11-dimethyl-8,9,lO,ll-tetrahydrobenz [alanthracene and characterized by means of its ultraviolet spectrum. This dimethylcholanthrene has a second methyl group in the same position that would be occupied by the steroid side chain. 2. Attempts to prepare 3-naphthoy1-5,7-&methylhydrindene, with the intention of pyrolyzing this ketone to 3,5-dimethylcholanthrene, were unsuccessful. Apparently the two ortho groups offer too much hindrance for the preparation of the ketone. 3. The Reformatsky reaction of S'-keto-3,4dihydro-l,2-~yclopentenophenanthrene with ethyl a-bromopropionate gave a product which could not be characterized. EVANSTON, ILLINOIS
RECEIVED AUGUST15, 1947
